Tolerance inerval vs. Distribution profiler

[#1]

Hi.

I've generated 20 numbers from standard normal distribution.

One sided tolerance interval for proportion of 97.5% is 2.114196.

On the othe hand, the distribution profiler shows that for 2.114196 the upper bound is 99.9959%

Am I missing something? I was thinking that the upper bound of the the distribution profiler has to be equal/approx. equal to the proportion of the tolerance interval.

I am trying to understand these numbers because of the following need.

I need to calc a tolerance interval for Gamma/Johnson distribution.

JMP does not have this option. I was thinking about distribution profiler of Gamma dist. as a solution.

Is there some way to cal a tolerance interval for Gamma/Johnson distribution?

Thank you!

[#2]

The distribution profiler shows the confidence interval, not the tolerance interval (at least it does that with my installation). This might explain why you are seeing 99.9959% "confidence" (not tolerance) for a value of 2.114196.